Update System: Not Rawhide

[Bill Nottingham]

Jason L Tibbitts III (tibbs at math.uh.edu) said: 
> >>>>> "JK" == Jeremy Katz <katzj at redhat.com> writes:
> 
> JK> If rawhide diverged earlier, we'd see a substantial drop in the
> JK> number of testers and thus the number of things found late in the
> JK> game.
> 
> I've seen that claim multiple times, but I'm not sure I've ever seen
> any evidence to back it up.  Even assuming that the claim is true,
> there's a need to balance those effects versus those of having the
> repo that's supposed to be stabilizing also be the one where all
> development has to happen.

How much testing is done by people who update to the latest rawhide
and try things? 

With koji, packages are available through the interface if people
want them. So we could set up builds for the next distribution
that people could still get if they wanted, all while having
rawhide for distribution testers pointing at the current in-development
tree.

Bill